Build a marketing intelligence dashboard

Overview

Computer can turn fragmented ad platform data into a unified, interactive marketing intelligence cockpit that performance teams can actually run the business from. It aggregates Meta, Google, TikTok, ASA, and MOLOCO into one dashboard with candlestick charts, pacing, forecasts, and one-click optimisation actions that are all logged for audit and learning

Key Output

Computer built and deployed the AdPulse marketing intelligence dashboard, a web app that surfaces five core KPIs (spend, CAC, subscriptions, signups, impressions) as headline cards with 7‑day sparklines and deltas, plus tabbed daily OHLC candlestick charts for each metric. Platform pills and segment/cohort filters drive dynamic updates across charts and KPI tiles so you can immediately see how efficiency shifts by channel or audience.

Channel contribution views combine stacked bars, a spend distribution donut, and a per‑platform table with ROAS, CAC, and trend sparklines to make relative performance obvious.

A budget pacing module shows overall and per‑channel progress against monthly targets with gauges, coloured progress bars, and a cumulative spend vs ideal line chart.

A predictions panel uses recent history (e.g., 7‑day moving averages) to generate next‑day spend and CAC forecasts by platform with confidence indicators, while a recommendations engine surfaces bullish scale suggestions and bearish cut actions as clickable cards. Approving an action triggers a logged entry capturing timestamp, what changed, and before/after metrics in a live action log table. The dashboard is responsive, structured to accept live API feeds (Google already partially connected), and ready to become the central control centre for cross‑channel budget decisions.

Expand

Expand

Expand

Tips

Clarify source of truth – Add guidance like: “Treat this dashboard as the single source of truth for CAC and ROAS; reconcile against platform-reported metrics weekly.”

Constrain recommendation logic – Specify risk bounds, for example: “Never recommend budget changes larger than ±25% day-over-day without a 7-day trend confirmation.”

Align with operating cadence – Tell Computer how you’ll run it, for example: “Optimise the layout for a 15-minute daily standup where we review pacing, forecasts, and approve/deny 3–5 recommended actions.”